© 2002 American Thoracic Society Lung-Specific Expression of Dominant-Negative Mutant p53 in Transgenic Mice Increases Spontaneous and Benzo(a)pyrene-Induced Lung CancerDepartments of Medicine, Microbiology, Pathology, and Environmental Medicine, Division of Pulmonary and Critical Care Medicine, and Department of Environmental Medicine, Division of Biostatistics, New York University School of Medicine, New York, New York Address correspondence to: Kam-Meng Tchou-Wong, Ph.D., New York University School of Medicine, 550 First Avenue, MSB 147, New York, NY 10016. E-mail: tchouk02{at}endeavor.med.nyu.edu
Mutations in the p53 gene have been implicated to play an important role in the development of various human cancers. To evaluate the importance of p53 in lung cancer, a transgenic mouse model was established by utilizing the Clara cell secretory protein (CCSP) promoter to target the expression of a dominant-negative mutant form of p53 (dnp53) in the lung. In two transgenic CCSP-dnp53 founder lines, the dnp53 protein was expressed exclusively in the lungs. The incidence of spontaneous lung cancer in 18-month-old transgenic mice was 45%, whereas that in age-matched control mice was 20%. The relative risk of lung tumors in CCSP-dnp53 mice was 2.3 times that of wild-type mice (exact confidence limits of 0.69, 17.5). In addition to the increased incidence of spontaneous lung tumor, these mice were more susceptible to the development of lung adenocarcinoma after exposure to benzo(a)pyrene (BaP). Six months after intratracheal instillation of benzo(a)pyrene, the tumor incidence in wild-type and CCSP-dnp53 mice was 39% and 73%, respectively. The risk of lung tumors was 25.3 times greater in BaP-treated mice adjusted for transgene expression (95% confidence limits of 3.29, 678, mid-p corrected). These results suggest that p53 function is important for protecting mice from both spontaneous and BaP-induced lung cancers.
Abbreviations: Clara cell secretory protein, CCSP • cytomegalovirus, CMV • dominant-negative mutant p53, dnp53 • ß-galactosidase, ß-gal • Li-Fraumeni syndrome, LFS • murine CCSP, mCCSP • polymerase chain reaction, PCR • retinoblastoma, Rb • T antigen, TAg
p53 is a tumor suppressor gene that is frequently mutated in lung cancer with mutations in > 50% of non–small cell lung cancer and 70% of small cell lung cancer (1, 2). Cigarette smoke exposure is correlated with G:C to T:A transversions in lung cancer, and p53 mutations in cigarette smokers occur at hotspots in contrast to the absence of mutation hotspots in nonsmokers. The tobacco carcinogen benzo(a)pyrene (BaP) forms DNA adducts on p53 preferentially at codons 157, 245, 248, and 273 in vitro in human bronchial epithelial cells (3). The DNA adducts occur more frequently at heavily methylated cytosines at CpG sites, and are repaired more poorly on the nontranscribed strand of DNA (4, 5). Whereas < 10% of smokers develop lung cancer, almost 50% of smokers with Li-Fraumeni syndrome (LFS), an autosomal dominant disorder characterized by germline mutations in the p53 gene, develop lung cancer. Because carriers of germline p53 mutations in LFS families are at increased risk for lung cancer, p53 plays a crucial role in the predisposition and development of smoking-related lung cancer (6). The p53 protein is a transcription factor that induces the expression of several genes, including p21, an inhibitor of the cyclin/cyclin-dependent kinase complex mediating G1 cell cycle arrest; bax, which promotes apoptosis; and GADD45, which is involved in cell cycle arrest and DNA nucleotide excision repair (7). The p53 nuclear protein contains 393 amino acids and has three important domains, including a transactivation domain at the amino terminal (residues 1–42), a DNA-binding domain (residues 102–292), and a carboxyl-terminal domain with a tetramerization motif (1). Most missense mutations occur in the DNA-binding domain, resulting in defective contacts with DNA and loss of the ability of p53 to act as a transcription factor. Chemically- and transgenically-induced primary lung tumors in mice share many morphologic, histogenic, and biochemical features of human adenocarcinomas. The reproducible natural history of these tumors allows molecular characterization at each stage of progression, specifically from premalignant to malignant lesions. Genetic factors influence the susceptibility to lung tumors in both mice and humans. Hence, inbred mice strains and their transgenic derivatives provide useful experimental tools for studying differences in susceptibility to lung tumorigenesis and sensitivity to carcinogens. Mice with disrupted germline p53 alleles, i.e., homozygous knockout mice, are more susceptible to the development of spontaneous tumors of various types, primarily lymphomas and sarcomas by 6 mo of age (8, 9). The p53 knockout mouse offers a tool for studying chemical carcinogenesis because heterozygous p53 knockout mice are more sensitive to carcinogens than normal mice (10). A second transgenic mouse containing a mutant p53val135 (alanine to valine change at codon 135 of the p53 gene) developed a high incidence of lung adenocarcinomas, osteosarcomas, and lymphomas (11, 12). Transgenic mice with the viral SV40 T antigen (TAg), which binds and inactivates both the p53 and the retinoblastoma (Rb) genes, driven by the Clara cell secretory protein (CCSP) promoter or the SP-C promoter developed a large number of lung adenocarcinomas by 6 mo (13–15). Because TAg inactivates both the p53 and Rb pathways, these mouse models did not address specifically the p53 pathway in the lung. The A/J mouse lung tumor model results in a high percentage of lung adenomas that harbor K-ras mutations, but p53 mutations in this model are uncommon (16–18). The C57BL/6, A/J, or C3H/HeJ strains exposed intratracheally to BaP plus charcoal developed squamous cell carcinomas at early time points at high doses (8 mg), followed by adenomas and adenocarcinomas at later time points and lower doses (4 mg) (19). Charcoal was necessary for the development of squamous cell carcinomas by impeding clearance of BaP. Zhang and colleagues combined the A/J mouse model with the p53+/- heterozygous or p53135 point mutation mice, resulting in increased adenomas with intraperitoneal carcinogen treatment, particularly in the A/J X p53135 mouse (6). We hypothesized that lung-specific expression of dominant-negative mutant p53 in transgenic mice will provide a mouse model for studying the role of p53 in both spontaneous and carcinogen-induced lung carcinogenesis. Although transgenic mice expressing TAg specifically in the lung under the CCSP or SP-C promoter have been shown to develop spontaneous lung adenocarcinomas (13–15), the susceptibility to carcinogen-induced lung tumorigenesis has not been examined. On the other hand, p53 knockout mice have been shown to be more susceptible to carcinogen-induced liver tumorigenesis (10). Because the contribution of Clara cells to the proliferation compartment of normal human tracheobronchial epithelium is substantial, Clara cells may play a role in the maintenance of the normal epithelium of the distal conducting airways in humans (20). For lung-specific expression of the mutant p53 gene, we chose to use the murine CCSP promoter because most non–small cell lung cancers are derived from bronchial epithelial cells, and the murine respiratory epithelium contains 50–60% Clara cells (21). To this end, we have generated transgenic mice expressing a dominant-negative mutant p53 (dnp53) under the 898-bp lung-specific CCSP promoter. This dnp53 mutant was generated by deletion of the DNA transactivation and sequence-specific DNA binding domains (amino acids 15–301) (22). Because the C-terminal oligomerization is intact, the dnp53 protein can oligomerize with wild-type p53. Hence, the dnp53 mutant may transform through a dominant-negative mechanism by generating DNA-binding–incompetent oligomers and inhibiting the DNA-binding activity of wild-type p53 (22).
Cloning and Characterization of Murine CCSP promoter Ray and coworkers showed that an 803 bp fragment of the 3-kb murine CCSP (mCCSP) promoter was active only in the lung and not in any other organs (23, 24). We cloned the mCCSP promoter from murine liver genomic DNA by polymerase chain reaction (PCR) using the high-fidelity enzyme pfu (Stratagene, La Jolla, CA). The mCCSP promoter fragment (898 bp) was inserted into the mammalian expression vector pcDNA3 (Invitrogen, San Diego, CA), replacing the human CMV promoter and enhancer, to generate the vector pmCCSP. To test the activity of the mCCSP promoter, ß-galactosidase (ß-gal) was inserted into the plasmid pmCC10–ß-gal and compared with ß-gal driven by the cytomegalovirus (CMV) promoter. After transient transfection into lung adenocarcinoma H441 cells (ATCC, Manassas, VA), pCMVß-gal increased ß-gal by 4.5-fold compared with 1.7-fold for pCCSP–ß-gal over the empty plasmid vector alone. These results were consistent with previous reports that the mCCSP promoter is less active in in vitro cell culture systems but possibly acquires full activity in vivo (23).
Expression of dnp53 by mCCSP Promoter in H441 Cells
Development of dnp53 Transgenic Mouse
Intratracheal Instillation
Evaluation of Transgenic Mouse Lungs
Statistical Analyses
Generation of Transgenic Mice and Lung-Specific Expression of dnp53 To generate CCSP-dnp53 transgenic mice, the dnp53 expression cassette fragment containing the mCCSP promoter, the dnp53 gene, and the poly A signal (Figure 1A) was injected into the pronucleus of fertilized eggs of FVB/N mice. The lungs of seven PCR-positive founder lines were screened by Western blot analysis using the PAb421 antibody directed against the C-terminus of p53 (amino acids 371–380). The expression of dnp53 protein was detected in two out of the seven founder lines (data not shown). The tissue specificity of transgene expression in the two positive lines 2,501 and 3,503 was examined by Western blot analysis. As shown in Figure 1B, the dnp53 protein was expressed exclusively in the lung, but not in the brain, heart, intestine, kidney, liver, muscle, spleen, or uterus (line 3,503). Hence, expression of the transgene under the CCSP promoter was lung-specific. Because the expression of dnp53 protein was higher in line 3,503, this line was used for all subsequent studies.
Incidence of Lung Cancer in CCSP-dnp53 Transgenic Mice When expressed in the choroid plexus epithelium, the dnp53 transgene induced brain tumor development, but expression of dnp53 in the liver had no adverse effect and the transgenic mice lived a normal life span (25). To determine the effects of expression of dnp53 in the lung, CCSP-dnp53 transgenic mice and wild-type mice were kept for 18 mo for the development of spontaneous lung cancer. As shown in Table 1 , the incidence of spontaneous lung cancer in FVB/N wild-type mice was 20% (4/20) and was increased to 45% (14/31) in CCSP-dnp53 mice. The exact probability of observing the results is 0.08 (Fisher's exact test, 2-sided). The relative risk of lung tumors in dnp53 mice is 2.3 times that of wild-type mice (exact 95% confidence limits of 0.69, 17.5).
Because these CCSP-dnp53 mice are at increased risk for the development of spontaneous lung cancer as compared with wild-type FVB/N mice (Table 1), suggesting that inhibition of p53 function by expression of dnp53 may predispose to the development of lung cancer, these transgenic mice are also useful for studying the effects of mutant p53 expression in susceptibility to carcinogen-induced lung cancer with a shortened latency. As a model of smoking-induced lung cancer, we exposed wild-type and CCSP-dnp53 mice to the tobacco carcinogen BaP via intratracheal instillation. Two-month-old mice were instilled with the tricaprylin solvent or BaP at weekly intervals over a period of 4 wk and killed six months later. Table 2 examines the effects of dnp53 expression in susceptibility to carcinogen-induced lung cancer. Because these mice were killed at an earlier age (9 mo), the incidence of spontaneous lung cancer in solvent-exposed wild-type mice and CCSP-dnp53 mice was 0% (0/8) and 10% (1/10), respectively. The incidence of lung tumors induced by BaP in wild-type mice and CC10-dnp53 mice were 39% (5/13) and 73% (8/11), respectively. The risk of lung tumors is 25.3 times greater in BaP-treated mice adjusted for transgene expression (95% confidence limits of 3.29, 678, mid-p corrected). The null hypothesis that the common relative risk is unity was rejected (P = 0.0006).
Immunohistochemical Staining of Normal Lung Mucosa and Lung Adenocarcinomas We performed immunohistochemical staining of lung sections with anti-CCSP polyclonal antibody and anti-p53 monoclonal antibody directed to the C-terminus of the p53 transgene. As shown in Figure 2A , expression of CCSP but not p53 was demonstrated in the normal mucosa (bronchus) of wild-type mice. On the other hand, expression of both CCSP and p53 was seen in the normal mucosa of CCSP-dnp53 mice (Figure 2B). Consistent with the specific expression of dnp53 protein in the lungs of CCSP-dnp53 mice by immunoblot analysis (Figure 1B), expression of dnp53 in the nucleus was demonstrated only in the bronchial epithelial cells lining the major airways (Figure 2B). Furthermore, immunohistochemical staining of the normal bronchus of a BaP-treated CCSP-dnp53 mouse with antibodies specific for p53, PCNA, GADD45, and p21 demonstrated the expression of p53 and PCNA in the nucleus, whereas expression of GADD45 and p21 was not detected (Figure 3) . All lung tumors, both spontaneous and BaP-induced, were adenocarcinomas as determined by H&E staining.
Although CCSP was expressed in the normal mucosa, tumors from both wild-type and CCSP-dnp53 mice were negative when stained with the anti-CCSP antibody (Figures 2A and 2B). However, these tumors stained positively for the transcription factor-1 (TTF-1) (data not shown), a major regulator of the CCSP promoter (24). Interestingly, these lung tumors also failed to stain with anti-p53 antibody, despite the intense p53 nuclear staining in the Clara cells of the normal bronchus of CCSP-dnp53 mice. Because the dnp53 transgene is expressed in normal progenitor cells but not in the established adenocarcinomas, it is of interest to determine when the dnp53 transgene is turned off during the carcinogenesis process. To examine if dnp53 was expressed at earlier stages of lung carcinogenesis, we examined the expression of p53 in an early neoplastic lesion (25 µ) and a large established tumor (1,000 µ). As shown in Figure 4A , expression of p53 was not detected in either the small lesion or large tumors, suggesting that CCSP-dnp53 was turned off even before the small tumor progressed to a larger size. In contrast, nuclear staining for GADD45 was detected only in the early lesion but not in the large tumor (Figure 4B), whereas nuclear PCNA staining was seen in both the established tumor and in the early lesion (Figure 4C).
The p53 tumor suppressor gene is critical in preventing cancer formation in humans and mice. The p53-deficient mice lacking one (p53+/-) or both (p53-/-) p53 alleles are more susceptible than their wild-type counterparts to the development of an array of different tumor types. Because of the enhanced tumor susceptibility of the p53+/- mice, this transgenic mouse model is being considered as a substitute for the standard 2-yr rodent carcinogenicity assays (26). It was hypothesized that p53 is haploinsufficient for tumor suppression and that reduction in p53 dosage is sufficient to promote cancer formation by providing an environment more conducive to the development of further oncogenic lesions and the initiation of tumorigenesis. To determine the role of p53 in lung carcinogenesis, we have generated a transgenic mouse model to investigate the consequences of targeted expression of a dominant-negative mutant p53 gene in the nonciliated bronchiolar epithelial Clara cells. In contrast to p53 knockout mice in which all cell types are affected, in the CCSP-dnp53 mice, mutant p53 is expressed only in the Clara cells. Hence, the CCSP-dnp53 mouse is generated to provide a mouse model for studying pulmonary adenocarcinoma. In the mouse, Clara cells are the principal cell type constituting the murine airway epithelium (21) and are the primary targets for metabolically activated pulmonary cytotoxicants and carcinogens (27). In addition, Clara cells are also the progenitor cells for the repair of the bronchiolar epithelium after injury and hence are critical for distal airway epithelial function and regeneration (28). Expression of mutant p53, which lacks the DNA-binding domain, will compromise wild-type p53 function by inhibiting p53-dependent transcription in a dominant-negative fashion. Because p53 may be a haploinsufficient tumor suppressor, the dominant-negative inhibitory effects of mutant p53 on wild-type p53 function may be analogous to the p53+/- mice, in which reduced p53 dosage is conducive to tumor initiation. Because the CCSP-dnp53 transgenic mice are more susceptible to both spontaneous and BaP-induced lung cancer, these mice should provide a useful model for carcinogenicity testing, especially for smoking-related lung carcinogens. Similar to two other transgenic mouse models of lung cancer induced by expression of the viral oncogene TAg under the CCSP or SP-C promoter (13, 15), the CC10-dnp53 mice also developed pulmonary adenocarcinomas, although with a longer latency period. The shorter latency for tumor development induced by TAg can be explained by the fact that TAg binds to and inhibits both p53 and Rb, thus inactivating two major tumor suppressor gene products. On the other hand, inhibition of the p53 pathway by expression of dnp53 can also promote the development of pulmonary adenocarcinomas. It has been demonstrated that pulmonary adenocarcinomas in humans express predominantly surfactant protein (29), and CCSP is infrequently expressed in human lung cancers, despite being expressed in progenitor cells for normal and neoplastic epithelium. This is consistent with the reduced and sporadic expression of the endogenous CCSP gene in lung adenocarcinomas of CCSP-TAg transgenic mice and in transformed Clara cells in vitro (13). Similarly, in SP-C–TAg mice, the expression of CCSP was consistently decreased in larger tumors at later stages of tumor progression (15). We have shown here that the expression of CCSP was also not detected in adenocarcinomas of wild-type and CCSP-dnp53 mice. In fact, overexpression of CC10 in A549 adenocarcinoma cells has been shown to reduce anchorage-independent growth and cell transformation (30), suggesting that expression of CCSP may be selected against in adenocarcinomas. Because CCSP antagonizes the transformed phenotype, downregulation of CCSP expression has been suggested to contribute to carcinogenesis (31). Hence, during the development of adenocarcinomas in these mice, the expression of CCSP may be downregulated due to selection against its expression. Alternatively, CCSP expression may reflect changes in gene expression associated with tumor progression or cell differentiation. In the multistep carcinogenesis of lung cancer, we have shown that PCNA, a marker of cell proliferation, was expressed in the normal bronchus, early lesion, and tumor of the CCSP-dnp53 mouse. However, GADD45 was not expressed in the normal bronchus nor in the tumor but was expressed in the early lesion, suggesting that GADD45 may be an early biomarker for lung cancer. In contrast, the dnp53 protein was expressed only in the normal bronchus but not in the early lesion nor the tumor. Because CCSP was not expressed in these tumors, it is conceivable that the CCSP gene is turned off and hence the dnp53 transgene, which is under the control of the CCSP promoter, is also turned off. Because p53 staining was not detected in the early lesion, expression of dnp53 transgene may be turned off early during tumorigenesis. To determine when the expression of CCSP and dnp53 is lost during tumor initiation and progression, further experiments will be needed in which animals will be killed at earlier time points to obtain preneoplastic and early lesions for immunohistochemical staining for CCSP and p53 expression. Our data suggest that expression of mutant p53 may be important for the initiation of the carcinogenesis process but is not required for the maintenance of transformation once the tumor is formed. Because mutant p53 competes with wild-type p53 and inhibits its function in a dominant-negative fashion, expression of dnp53 may be analogous to the acquisition of loss-of-function mutations in p53 in human cancers. Both loss-of-function and gain-of-function mutations have been described for the p53 gene (32). Although mutations in the p53 gene are not common in murine lung tumors (33), the lack of p53 protein expression in the tumors of both wild-type and CCSP-dnp53 mice suggests that loss or inhibition of p53 function is sufficient for the initiation of transformation of lung cancer in murine models. Because the incidence of adenocarcinomas is increasing in the United States (34), the increased susceptibility of the CCSP-dnp53 transgenic mice to the development of pulmonary adenocarcinomas makes this a useful mouse model for lung cancer. The CCSP-dnp53 mice will also be valuable for mechanistic studies underlying other known tobacco-specific carcinogens such as 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone and for the screening for other potential pulmonary carcinogens. These mice will also be useful for testing for the efficacy of chemopreventive or chemotherapeutic agents in tobacco-induced lung cancer.
The authors thank Dr. Arnold J. Levine for helpful advice and discussions. They also thank Drs. Moshe Oren and Francesco J. DeMayo for dominant-negative p53 mutant and anti-CCSP antibody, respectively. They also thank Dr. Chuanxiang Chi for assistance with animal work. This work was supported by Lola and Allen Goldring Clinical Scholars Fund (to Y.J.); and by National Institutes of Health Grants MO1-RR00096, RO1-HL59832, 62055 (to W.N.R.), RO1-ES09161 (to K.-M.T.-W.), and P30-ES00260 (NIEHS Center of Excellence Grant). Received in original form December 13, 2001 Received in final form March 26, 2002
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
